[tune] Add Nevergrad to Tune (#3985)

python/ray/tune/examples/nevergrad_example.py

@@ -0,0 +1,56 @@
+"""This test checks that Nevergrad is functional.
+
+It also checks that it is usable with a separate scheduler.
+"""
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+
+import ray
+from ray.tune import run_experiments, register_trainable
+from ray.tune.schedulers import AsyncHyperBandScheduler
+from ray.tune.suggest import NevergradSearch
+
+
+def easy_objective(config, reporter):
+    import time
+    time.sleep(0.2)
+    for i in range(config["iterations"]):
+        reporter(
+            timesteps_total=i,
+            neg_mean_loss=-(config["height"] - 14)**2 +
+            abs(config["width"] - 3))
+        time.sleep(0.02)
+
+
+if __name__ == '__main__':
+    import argparse
+    from nevergrad.optimization import optimizerlib
+
+    parser = argparse.ArgumentParser()
+    parser.add_argument(
+        "--smoke-test", action="store_true", help="Finish quickly for testing")
+    args, _ = parser.parse_known_args()
+    ray.init(redirect_output=True)
+
+    register_trainable("exp", easy_objective)
+
+    config = {
+        "nevergrad": {
+            "run": "exp",
+            "num_samples": 10 if args.smoke_test else 50,
+            "config": {
+                "iterations": 100,
+            },
+            "stop": {
+                "timesteps_total": 100
+            },
+        }
+    }
+    optimizer = optimizerlib.OnePlusOne(dimension=2)
+    algo = NevergradSearch(
+        optimizer, ["height", "width"],
+        max_concurrent=4,
+        reward_attr="neg_mean_loss")
+    scheduler = AsyncHyperBandScheduler(reward_attr="neg_mean_loss")
+    run_experiments(config, search_alg=algo, scheduler=scheduler)

python/ray/tune/suggest/__init__.py

@@ -3,6 +3,7 @@ from ray.tune.suggest.basic_variant import BasicVariantGenerator
 from ray.tune.suggest.suggestion import SuggestionAlgorithm
 from ray.tune.suggest.bayesopt import BayesOptSearch
 from ray.tune.suggest.hyperopt import HyperOptSearch
+from ray.tune.suggest.nevergrad import NevergradSearch
 from ray.tune.suggest.skopt import SkOptSearch
 from ray.tune.suggest.sigopt import SigOptSearch
 from ray.tune.suggest.variant_generator import grid_search, function, \
@@ -13,6 +14,7 @@ __all__ = [
     "BasicVariantGenerator",
     "BayesOptSearch",
     "HyperOptSearch",
+    "NevergradSearch",
     "SkOptSearch",
     "SigOptSearch",
     "SuggestionAlgorithm",

python/ray/tune/suggest/nevergrad.py

@@ -0,0 +1,89 @@
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+
+try:
+    import nevergrad
+except Exception:
+    nevergrad = None
+
+from ray.tune.suggest.suggestion import SuggestionAlgorithm
+
+
+class NevergradSearch(SuggestionAlgorithm):
+    """A wrapper around Nevergrad to provide trial suggestions.
+
+    Requires Nevergrad to be installed.
+    Nevergrad is an open source tool from Facebook for derivative free
+    optimization of parameters and/or hyperparameters. It features a wide
+    range of optimizers in a standard ask and tell interface. More information
+    can be found at https://github.com/facebookresearch/nevergrad.
+
+    Parameters:
+        optimizer (nevergrad.optimization.Optimizer): Optimizer provided
+            from Nevergrad.
+        parameter_names (list): List of parameter names. Should match
+            the dimension of the optimizer output.
+        max_concurrent (int): Number of maximum concurrent trials. Defaults
+            to 10.
+        reward_attr (str): The training result objective value attribute.
+            This refers to an increasing value.
+
+    Example:
+        >>> from nevergrad.optimization import optimizerlib
+        >>> optimizer = optimizerlib.OnePlusOne(dimension=1, budget=100)
+        >>> config = {
+        >>>     "my_exp": {
+        >>>         "run": "exp",
+        >>>         "num_samples": 10,
+        >>>         "stop": {
+        >>>             "training_iteration": 100
+        >>>         },
+        >>>     }
+        >>> }
+        >>> algo = NevergradSearch(
+        >>>     optimizer, max_concurrent=4, reward_attr="neg_mean_loss")
+    """
+
+    def __init__(self,
+                 optimizer,
+                 parameter_names,
+                 max_concurrent=10,
+                 reward_attr="episode_reward_mean",
+                 **kwargs):
+        assert nevergrad is not None, "Nevergrad must be installed!"
+        assert type(max_concurrent) is int and max_concurrent > 0
+        self._max_concurrent = max_concurrent
+        self._parameters = parameter_names
+        self._reward_attr = reward_attr
+        self._nevergrad_opt = optimizer
+        self._live_trial_mapping = {}
+        super(NevergradSearch, self).__init__(**kwargs)
+
+    def _suggest(self, trial_id):
+        if self._num_live_trials() >= self._max_concurrent:
+            return None
+        suggested_config = self._nevergrad_opt.ask()
+        self._live_trial_mapping[trial_id] = suggested_config
+        return dict(zip(self._parameters, suggested_config))
+
+    def on_trial_result(self, trial_id, result):
+        pass
+
+    def on_trial_complete(self,
+                          trial_id,
+                          result=None,
+                          error=False,
+                          early_terminated=False):
+        """Passes the result to Nevergrad unless early terminated or errored.
+
+        The result is internally negated when interacting with Nevergrad
+        so that Nevergrad Optimizers can "maximize" this value,
+        as it minimizes on default.
+        """
+        ng_trial_info = self._live_trial_mapping.pop(trial_id)
+        if result:
+            self._nevergrad_opt.tell(ng_trial_info, -result[self._reward_attr])
+
+    def _num_live_trials(self):
+        return len(self._live_trial_mapping)